fax: (49) 40-42803-5359
Maid (GCIP) is involved in cell cycle control of hepatocytes†
Article first published online: 26 JAN 2007
Copyright © 2007 American Association for the Study of Liver Diseases
Volume 45, Issue 2, pages 404–411, February 2007
How to Cite
Sonnenberg-Riethmacher, E., Wüstefeld, T., Miehe, M., Trautwein, C. and Riethmacher, D. (2007), Maid (GCIP) is involved in cell cycle control of hepatocytes. Hepatology, 45: 404–411. doi: 10.1002/hep.21461
Potential conflict of interest: Nothing to report.
- Issue published online: 26 JAN 2007
- Article first published online: 26 JAN 2007
- Manuscript Accepted: 11 OCT 2006
- Manuscript Received: 12 JUL 2006
- Deutsche Forschungsgemeinschaft. Grant Number: SFB 444
- DFG. Grant Number: Li 1045/2-1
The function of Maid (GCIP), a cyclinD-binding helix-loop-helix protein, was analyzed by targeted disruption in mice. We show that Maid function is not required for normal embryonic development. However, older Maid-deficient mice—in contrast to wild-type controls—develop hepatocellular carcinomas. Therefore, we studied the role of Maid during cell cycle progression after partial hepatectomy (PH). Lack of Maid expression after PH was associated with a delay in G1/S-phase progression as evidenced by delayed cyclinA expression and DNA replication in Maid-deficient mice. However, at later time points liver mass was restored normally. Conclusion: These results indicate that Maid is involved in G1/S-phase progression of hepatocytes, which in older animals is associated with the development of liver tumors. (HEPATOLOGY 2007;45:404–411.)
Maid was originally isolated as a maternally transcribed helix-loop-helix (HLH) protein.1 Additionally, it was detected by screening for HLH proteins in fetal liver and called human homologue of Maid.2 Later it was independently isolated as a cyclinD interacting protein and named GCIP or DIP1.3, 4
Maid was shown in vitro to act as an inhibitory HLH protein, for example, blocking transcription of the HNF-4 promoter.2 In its function as a cyclinD1-binding protein it is able to reduce CDK4-mediated phosphorylation of the retinoblastoma protein and inhibit E2F-mediated transcriptional activity.3 Furthermore it is able to influence transcription of cyclinD1 and act as a tumor suppressor in the liver.5 In other recent reports human homologue of Maid (GCIP) was shown to also interact with p29 and Jun activation domain-binding protein 1, which both are implicated in cell cycle control.6, 7 From these in vitro data, Maid appears to function during differentiation or cell cycle control. Maid is expressed in a wide variety of adult tissues such as brain, intestine, muscle, and heart as well as in the unfertilized egg and the early embryo.1–3 After partial hepatectomy (PH), higher Maid expression was observed 24 to 36 hours after surgery.2
PH is a well-established method to examine the regenerative capacity of the liver.8 The adult liver retains its capacity to restore organ mass in response to liver injury such as PH, liver transplantation, or toxic liver cell damage. After 70% PH, approximately 95% of the remnant normally quiescent hepatocytes enter cell cycle to restore the original liver mass within 7 to 10 days.9, 10 Hepatocytes synchronously exit G0 and re-enter the cell cycle. As a consequence, 48 hours after PH, 30% to 50% of the hepatocytes are in S-phase.
Earlier studies defined two steps that are essential for cell cycle progression of hepatocytes. The first step is called the ”priming” phase, in which hepatocytes leave G0 and enter G1, where immediate early genes (transcription factors such as c-jun, c-myc, and c-fos) are expressed.9, 11 Further progression through the cell cycle is regulated by temporal activation of multiple cyclin-dependent kinases (CDKs). The formation of several CDK-cyclin-complexes is critical for progression through S-phase as cyclinE/A-CDK2 cooperates with cyclinD-CDK4/6 to phosphorylate the retinoblastoma (RB) protein, which releases bound E2F transcription factor. As a consequence, genes that are directly involved in the control of hepatocyte proliferation are activated.12, 13
To investigate the in vivo role of Maid, we generated a null mutation of Maid in mice. In these mice the functional role of Maid was investigated during embryonic development, in aging livers, and during liver regeneration after PH.
Materials and Methods
Construction of Targeting Vector and Generation of Maid Knockout Animals.
The targeting vector was created using standard techniques (detailed information available on request). The linearized targeting vector was introduced by electroporation into embryonic stem (ES) cells. After homologous recombination and Cre-mediated removal of exon 4 and neo-cassette, the ES cells were injected into C57BL/6 blastocysts and gave rise to germline chimeras. Heterozygous and homozygous mice were identified by PCR (primer sequences available on request) and verified by Southern blot. All animals received humane care according to the criteria outlined in the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication 86-23 revised 1985).
In Situ Hybridization and Immunohistochemistry.
Embryos were isolated at E10-E18 from staged pregnancies and fixed in 1% paraformaldehyde (PFA) for several hours up to overnight. Thereafter samples were equilibrated in 20% sucrose, embedded in QCT (Miles), frozen, and cryosected. In situ hybridization was performed as described.14 As a Maid-specific probe, the complete cDNA was used. Livers were isolated 24 hours to 21 days after PH, immediately embedded in QCT (Miles), frozen, and cryosected. Immunohistochemistry was performed essentially as described by Brockschnieder et al.15 Antibodies used were the following: anti-cyclinA (sc-751, Santa Cruz Biotechnology), anti-cyclinE (sc-481, Santa Cruz Biotechnology). In the BrdU-labeling experiments anti-BrdU antibodies (Roche) were used to label BrdU-positive cells. DAPI staining was used to visualize the absolute number of cells in the liver.
Two-Thirds Hepatectomy and Isolation of Livers.
Six- to nine-week-old male Maid−/− and +/+ (control) animals were used for the experiments. Two-third PH was performed as described earlier.16 Animals were sacrificed 24 hours, 40 hours, 48 hours, 60 hours, 72 hours, 96 hours, 7 days, or 21 days after PH. For each time point indicated, at least three animals per genotype were used. Livers were isolated, weighed, and either shock frozen or fixed in 4% PFA. The genotype of the mice was additionally confirmed by PCR at the end of the experiments.
Western Blot Analysis.
Western blots, immunoprecipitation, or kinase assays were performed as described,17 using antibodies against cyclins D1, A, or E (sc-8396, sc-751, and sc-481; Santa Cruz), cdk2 (Ab-3; Neomarkers), cdk4, RB (ACD1, G3-245; BD Biosciences), phosphorylation of the retinoblastoma protein (pSpT249/252; Biosource) as primary antibodies and peroxidase-conjugated goat anti-rabbit, -rat and -mouse IgG (Jackson) as secondary antibodies. The antigen–antibody complexes were visualized using the ECL detection kit as recommended by the manufacturer (Amersham).
Ten to 166 hours after partial hepatectomy and 2 hours before killing, 30 μg BrdU per gram mouse were injected intraperitoneally. To label more proliferating cells, animals were injected with BrdU on 2 consecutive days and were given BrdU in their drinking water (1 mg/ml) for 2 or 6 days, respectively.
For BrdU labeling of embryos, 30 μg BrdU per gram body weight was injected intraperitoneally into the pregnant female 1 hour before the animal was killed.
Northern Blot and Reverse Transcription PCR Analysis.
Northern blots were performed on 30 μg total RNA using 32P-dCTP-labeled full-length Maid and GAPDH as probes. For reverse transcription PCR analysis, mRNA from livers of wild-type and homozygous mutant mice was prepared before and after PH; cDNA was prepared from the RNAs and subjected to different rounds of PCR cycles with primers specific for Maid: 5′ primer: gcggcctcttgaatgattctg, 3′ primer: agctgggccacctcatcctt, or GAPDH: 5′ primer: accacagtccatgccatcac, 3′ primer: tccaccaccctgttgctgta. After 24, 27, 30, 33, and 36 cycles, probes were taken and analyzed on agarose gels.
Maid in Embryonic Development.
Former analysis of Maid expression had concentrated on adult tissue and the zygote. Presence of maid transcripts during embryonic development had only been demonstrated by Northern blot analysis at midgestation (E7.5 and E9).1, 2 To obtain more accurate information on the expression of Maid during embryonic development, we performed a detailed expression analysis by in situ hybridization. During organogenesis (E12) expression of Maid could be found in liver, the central nervous system, and dorsal root ganglia (Fig. 1A). This expression persisted throughout embryonic development. After E13 additional organs showed positive Maid signals, including kidney, intestine, nasal cavities, lung, and thymus (Fig. 1B).
Generation and Embryonic Analysis of Homozygous Maid Mutant Mice.
To analyze the role of Maid in vivo, we generated mice carrying a mutation in the Maid gene by homologous recombination. We constructed a vector in which exon 4 of Maid is flanked by two loxP-sites and an additional loxP-site is present downstream of the neomycin resistance cassette (Fig. 2A).
The vector was transfected into E14 ES-cells. ES-cells, where homologous recombination had occurred, were identified by Southern blot analysis (Fig. 2B) and subsequently transfected with a Cre-expression vector. Cre activity resulted in a deletion of Exon 4 and the neo-cassette (Fig 2A,B). Because exon 4 consists of 82 bp, deletion of this exon induces a frameshift, thereby leading to a complete inactivation of maid.
These ES cells were used to create mice carrying a mutated allele of Maid. Heterozygous mutant animals were crossed, and the offspring of those matings showed the expected 1:2:1 ratio of wild-type to heterozygous to homozygous animals. This result indicated that Maid was not essential for normal embryonic development. Animals were genotyped by PCR (Fig. 2C) and verified by Southern blot analysis to prove the reliability of the PCR (data not shown).
To confirm the inactivation of Maid on the RNA level, Northern blot analysis was performed on RNA derived from wild-type, heterozygous, and homozygous mutant embryonic liver using a Maid-specific probe (complete cDNA) (Fig. 2D). In contrast to RNA from wild-type and heterozygous mutant embryos, no Maid-specific RNA could be detected in Maid−/− embryos, demonstrating that mutant, non-functional RNA is rapidly degraded and the generated allele is a true knockout. Rehybridization of the filter with a GAPDH-specific probe showed that similar amounts of RNA were loaded (Fig. 2D).
According to the widespread expression of Maid in adult tissues, we investigated numerous aspects of postnatal life of Maid-deficient animals. No differences in comparison with wild-type animals were found in morphology, histology, fertility, motility, and weight. We did not observe an increased mortality during the first 12 months. Because the expression in the nervous system is comparatively high, we also investigated nociception and odor-induced variation in anxiety-like behavior, but again found no differences. We thus concluded that Maid-deficient animals are viable and not impaired when kept under standard conditions.
Lack of Maid Expression Results in the Earlier Development of Liver Tumors.
Because lack of Maid expression had no effects on embryonic development and during the first year of adult life, we investigated older animals. Interestingly, in Maid-deficient male animals from 1 to 1.5 years of age, 4 of 32 animals (12.5%) developed liver tumors (Fig. 3A,B). In contrast, no liver tumors were found in male wild-type mice (0 of 28) and female wild-type and Maid−/− animals (Fig. 3A and data not shown). Further histological analysis classified these tumors as hepatocellular adenomas (HCAs) and hepatocellular carcinomas (HCCs) (Fig. 3C,D). As it is known from the literature, murine as well as human females develop liver tumors later in life compared with males.18, 19 We therefore monitored tumor formation in females aged 1.5 to 2 years. No tumors were found in female wild-type mice (0/22), whereas in female Maid mutant mice 7 of 22 animals (32%) developed HCAs and HCCs. These results indicated that lack of Maid expression in the liver triggers earlier tumor development, whereas in no other tissue an increase in tumor rates was observed (data not shown).
Maid Expression in the Liver.
Because liver tumors occurred earlier in male and female Maid−/− mice, our further experiments concentrated on the physiological role of Maid in the liver. Histological sections of the embryonic liver showed no significant differences between wild-type and homozygous mutant embryos (Fig. 4). Because Maid had been shown to interact with cyclinD1 in vitro, we specifically investigated proliferation rates in embryonic livers between E10 and E16. However, no differences in BrdU-uptake were detected in livers derived from wild-type and Maid−/− embryos.
Next, we investigated the livers of adult animals. Up to 9 months of age, we could not detect any differences between livers of wild-type and homozygous mutant animals in histological sections (Fig. 4), and additionally the liver-to-body-weight ratio was not changed. The proliferation rate of the adult liver, which is very low in wild-type mice, was also not altered in homozygous mutant animals (data not shown).
Maid Expression During Liver Regeneration.
Because earlier experiments showed increased expression of Maid after PH,2 we performed PH experiments using wild-type and Maid-deficient male mice. In agreement with published results, we detected an increase in Maid mRNA expression 40 hours after PH in wild-type mice whereas, as expected, no Maid mRNA expression could be detected in Maid−/− animals. Sham-treated mice under these conditions showed no detectable Maid mRNA, indicating that Maid expression in resting adult hepatocytes is low (Fig. 5).
In further experiments, we tested the impact of a lack in Maid expression on cell cycle progression in hepatocytes. In wild-type mice, a maximum in DNA synthesis—as evidenced by BrdU staining—was found 40 to 48 hours after PH (Fig. 6A,B). In contrast, no increase in BrdU-positive cells could be detected in homozygous mutant mice 40 to 48 hours after PH. Instead, a delayed and less prominent peak could be observed in those animals at 60 hours after PH (Fig. 6B). Consistent with this delay in proliferation the liver-to-body-mass index was reduced in mutants compared with wild-type 72 hours after PH (Fig. 6D).
These results were further confirmed by long-term BrdU injection experiments. When BrdU was given over the first 48 hours after PH in wild-type mice, 40% of hepatocytes were positive. In contrast, only 8% of the liver cells were labeled in the Maid−/− animals (Fig. 6C). However, when mice were injected on the second and third or from the first to the 6th day after PH, no significant differences in the numbers of BrdU-positive cells in the liver were evident between wild-type and homozygous mutant (Fig. 6C), showing that the total proliferation over time is comparable. These results indicate that Maid is involved in triggering G1/S-phase transition of hepatocytes during cell cycle progression.
Expression Analysis of Cyclins, CDKs, and RB After PH.
BrdU analysis revealed that lack of Maid expression delays the start of DNA synthesis in hepatocytes after PH. To better understand the mechanism of how Maid influences cell cycle progression in resting hepatocytes, we performed additional experiments. CyclinD and E expression levels showed no strong difference between wild-type and Maid−/− animals after PH in Western blot analyses (Fig. 7A). To verify equal loading of the gels, blots were consecutively probed with a GAPDH-specific antibody (Fig. 7A).
In contrast, clear differences in cyclinA expression were found comparing wild-type and Maid−/− mice. In wild-type mice, strong cyclinA expression was first detected 48 hours after PH, whereas in livers of Maid−/− mice, an increase was observed only 60 hours after PH and seemed not as robust (Fig. 7A). Interestingly, the second peak of cyclinA expression in controls at 72 hours was not evident in mutant liver tissue where only one prolonged peak from 60 to 72 hours was present. The delay in cyclinA expression is in good agreement with the BrdU analysis and indicates that Maid is required in mediating G1/S-phase transition of hepatocytes during cell cycle progression.
To characterize this defect in greater detail, we monitored the phosphorylation state of RB and again found a significant difference between wild-type and mutant animals 48 hours after PH (Fig. 7B). Kinase experiments with immunoprecipitated CDK2 are in agreement with these data and showed a reduced activity in mutants at 48 hours after PH (Fig. 7C). Interestingly, we found no significant discrepancies between wild-type and mutant when analyzing the formation of cyclinD-CDK4, cyclinE-CDK2, and cyclinA-CDK2 complexes 48 hours after PH (Fig. 7D).
Maid in Embryonic Development and Differentiation.
Different functions have been attributed to the Maid protein when its function was analyzed in vitro. It has been shown to work as an Id-like-molecule1 as well as being able to bind to cyclinD1.3 To identify its functions in vivo, we generated a mouse strain carrying a mutated Maid gene in a homozygous form using standard protocols. By Northern blot analysis, we confirmed that homozygous animals are true null-alleles for Maid. Surprisingly, these animals were viable and showed a normal morphology in all tissues analyzed. Despite the strong expression in several tissues, Maid does not have essential functions during embryonic development. Especially the function as an Id-like molecule interfering with basic helix-loop-helix protein mediated transcription demonstrated by in vitro experiments2 and the high expression in the embryonic liver suggested a potential involvement in liver development. However, the liver develops completely normally in the absence of Maid, although HNF-4, whose expression can be regulated by Maid in vitro,2 plays an important role in the terminal differentiation of hepatocytes.20
The binding to cyclinD1 as demonstrated by in vitro experiments3 implicated Maid in the control of cell cycle regulation. By BrdU analysis, we were, however, unable to find significant differences in proliferation rates during embryonic development between wild-type and homozygous mutant mice focusing on neural and hepatic tissues. At all embryonic stages analyzed, central nervous system, peripheral nervous system, and liver displayed similar morphology and size when comparing wild-type and Maid-deficient embryos. This does not formally rule out that Maid is able to influence proliferation and differentiation, but it does show that the presence of Maid is dispensable for normal embryonic development.
Maid in Liver Regeneration.
Maid is expressed throughout embryonic liver development, whereas postnatally expression levels decline to very low levels.20 In good agreement with these low expression levels, Maid deficiency did not affect liver homeostasis and postnatal appearance of the liver during the first year. The finding that Maid is upregulated after PH2 prompted us to investigate liver regeneration in Maid-deficient animals. After performing PH, we could observe a delay in the proliferative response in homozygous mutants compared with wild-type mice. Additionally the sharp proliferation peak characteristic for wild-type mice could not be observed in homozygous mutant animals. These results clearly demonstrated that Maid is involved in liver regeneration after PH. However, mutant animals as wild-type controls were able to regenerate their liver completely during 8 days following PH. This indicated that Maid is not absolutely essential for liver regeneration, but is important for the synchronous and quick cell cycle entry, because in the absence of Maid synchronous cell cycle entry after 48 hours is strongly disturbed.
To further characterize the defect, we investigated some of the known in vitro effects of Maid in liver tissue after PH. Maid has been shown to interact with cyclinD.3 D-cyclins represent a unique component of the cell cycle apparatus as they serve as links between the extracellular environment and the core cell cycle machinery.21 During cell cycle reentry, cyclinD-CDK complexes drive the phosphorylation of the retinoblastoma gene, resulting in a release of the phosphorylation of the retinoblastoma protein–bound E2F transcription factors and consequently activation of E2F-target genes, including cyclinE.21 According to the in vitro results in which Maid (human homologue of Maid/GCIP) negatively influenced cyclinD-dependent CDK4 kinase activity and E2F-mediated transcriptional activity, lack of Maid should thus result in precocious cell cycle entry and earlier activation of cyclinE.3 Interestingly, in our experiments we could not detect changes in the timing of cyclinE expression levels comparing wild-type and Maid-deficient animals after PH, arguing against a direct involvement of Maid in E2F-mediated transcription in this context. We had the impression that cyclinD levels were elevated in mutant livers compared with control livers. This is in good accordance with recent findings that the overexpression of Maid leads to a reduction of cyclinD.5 However, the activity of cyclinD–CDK4 complexes seemed unaltered when wild-type and mutant liver extracts were analyzed by in vitro kinase assay (data not shown).
We have shown that Maid deficiency leads to a delayed expression of cyclinA, a reduction of RB phosphorylation, and thus to a delayed G1 to S transition. This negative influence on cell cycle progression in the absence of Maid is in agreement with results from another report that showed that overexpression of human Maid in NHDF and HepG2 cells led to an increase in the number of cells in S-phase.7
Our data suggest that the induction of maid expression after PH facilitates entry of the quiescent hepatocytes into the cell cycle. Possibly this effect is mediated via an interaction of Maid and cyclinD. Apart from phosphorylation, the cyclinD-CDK complexes also serve to titrate the cell cycle inhibitors p27Kip1 and p21Cip1 away from cyclinE- and A-CDK complexes,21 and Maid could be involved in this process. Our data support this theory because we could show a reduction of cdk2 activity in the absence of Maid.
Surprisingly, the activation of cyclinE and A complexes is not abolished in cells lacking cyclinD, revealing the presence of additional, cyclinD-independent mechanisms that link mitogenic stimuli with the activation of cyclinE- and A-CDK complexes.17 In the light of these open questions, it will be interesting to determine the exact mechanism by which the lack of maid expression results in delayed expression of cyclinA and reduced CDK2 activities and RB phosphorylation.
Maid in Tumor Formation.
A variety of mouse mutants or transgenic animals exist that show an increase in liver tumor formation,22, 23 among them mice overexpressing cyclinD1 in the liver.24 In such mice, the expression level of cyclinD1 correlates with the onset of HCCs. A recent study showed that, apart from cyclinD, which was found to be overexpressed in one third of human HCCs, cyclinE overexpression contributes to the development of HCCs, whereas downregulation of p27Kip, which was found in biologically aggressive HCCs, seems to be more important for tumor progression.25 As with other cancers, HCC is caused by the accumulation of several genetic alterations that finally result in unrestrained growth of normally quiescent hepatocytes. The absence of Maid led to an increase in liver tumor formation in male mice older than 12 months and in females older than 18 months. The earlier occurrence of liver tumors in male compared with female mice is well established and has been observed in several mouse mutants as well as in humans.23, 26 The onset of tumor occurrence in Maid-deficient animals was rather late compared with the other mouse models,22–24 which might reflect the relatively low expression levels of cyclinD and Maid in young hepatocytes. Probably other alterations (e.g., increasing of cyclinD levels) are required before the modulating activity of Maid becomes important. In this context, it is important to note that overexpression of GCIP leads to a reduced expression of cyclinD,5 and thus the absence of Maid might directly account for an increase of cyclinD levels in our model. Additionally, Maid (GCIP) has been shown in vitro to negatively affect cyclinD-dependent CDK4 kinase activity.3 All these data indicate that the absence of Maid can contribute to a higher tumor incidence in the liver. Along this line, another study has demonstrated interactions of Maid (GCIP) with p29,6 which also might be involved in cell cycle regulation. Interestingly, a recent study has shown that Maid (GCIP) can interact with Jun activation domain-binding protein 1,7 which is known to interact with several proteins, including p27Kip.27, 28 In this study, the authors also found that Maid overexpression is associated with hepatocarcinogenesis in humans and suggest that this effect is mediated by Maid–Jun activation domain-binding protein 1 interaction.7 In contrast, our data show that lack of Maid promotes the occurrence of HCCs in mice, and data from Ma et al.5 demonstrate that overexpression of GCIP suppresses liver tumorigenesis. Our results and results from the literature give conflicting data: Lack of Maid leads to a delay in proliferation (after PH) as well as to enhanced tumor formation (in aged livers). Similarly, overexpression of Maid resulted in reduced proliferation3 and tumor formation5 as well as enhanced proliferation.7 All of these data indicate that, depending on the environment and the expression level, Maid has a modulating function during cell cycle progression and is involved in liver tumor formation.
We thank Lynn Ellenberg and Yvonne Pechmann for expert technical assistance. We also want to thank Nils-Holger Zschemisch for help in some of the PH experiments.